[Kernel] DeepGemm MoE : Integrate triton permute / unpermute kernels (#20903)

Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com> Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
2025-07-17 13:40:37 +05:30
parent fdc5b43d20
commit 11dfdf21bf
10 changed files with 490 additions and 58 deletions
--- a/tests/kernels/moe/modular_kernel_tools/cli_args.py
+++ b/tests/kernels/moe/modular_kernel_tools/cli_args.py
@@ -85,7 +85,6 @@ def make_config_arg_parser(description: str):
                        help="num topk")
    parser.add_argument(
        "--fused-moe-chunk-size",
-        nargs="+",
        type=int,
        help="Fused moe chunk size used for the non-batched fused experts impl."
    )